2,979 publications from this institution
The highly diversified conceptual and algorithmic landscape of Granular Computing calls for the formation of sound fundamentals of the discipline, which cut across the diversity of formal frameworks (fuzzy sets, sets, rough sets) in which information granules are formed and processed. The study addresses this quest by introducing an idea of granular models – generalizations of numeric models that are formed as a result of an optimal allocation (distribution) of information granularity. Information granularity is regarded as a crucial design asset, which helps establish a better rapport of the resulting granular model with the system under modeling. A suite of modeling situations is elaborated on; they offer convincing examples behind the emergence of granular models. Pertinent problems showing how information granularity is distributed throughout the parameters of numeric functions (and resulting in granular mappings) are formulated as optimization tasks. A set of associated information granularity distribution protocols is discussed. We also provide a number of illustrative examples.
Over the recent years, we have been witnessing spectacular achievements of Machine Learning with highly visible accomplishments encountered, in particular, in natural language processing and computer vision impacting numerous areas of human endeavours. Driven inherently by the technologically advanced learning and architectural developments, Machine Learning constructs are highly impactful coming with far reaching consequences; just to mention autonomous vehicles, health care imaging, decision-making in critical areas, among others.
In this section, desirable features of fuzzy-logic-based hardware are discussed from the viewpoint of fuzzy logic as well as semiconductor design and implementation. We first discuss the need for fuzzy hardware and then provide some insight into the features that are important for hardware design and explain the reasons for these features.
Most fuzzy models are just numeric. In this study, we revisit, explore and augment a concept of linguistic models, viz., fuzzy models producing results that are information granules, and, specifically, intervals or fuzzy sets. The proposed architecture is formed by constructing a network of linked fuzzy sets (information granules) ininput and output spaces with the aid of a context-based Fuzzy C-Means clustering method. The user centricity of such clustering method is implied by the explicit formulation of fuzzy sets in the output space. The resulting information granules constructed in the input space are conditioned by the corresponding fuzzy sets in the output space. This arrangement can increase the interpretability of the model and represent the model as a collection of logically arranged associations among information granules. The model's overall design process is discussed along with a detailed algorithmic structure. Its experimental evaluations are provided by using both synthetic and publicly datasets. For the former, the model brings the performance improvement ranging from 91% to 250% over the models with information granules uniformly distributed in output space. For the latter, such improvement ranges from 6% to 94%. Finally, a thorough discussion is provided together with guidelines on how to develop such a linguistic model in different contexts.
The paper presents an analog, current-mode circuit that cal- culates a distance between the neuron weights vectors W and the input learning patterns X. The circuit can be used as a component of dierent self-organizing neural networks (NN) implemented in the CMOS technol- ogy. In Self-Organizing Maps (SOM) as well as in NNs using the Neural Gas or the Winner Takes All (WTA) learning algorithms, to calculate the distance between X and W , the same circuit can be used that makes it a universal structure. Detailed system level simulations of the WTA NN and the Kohonen SOM showed that using both the Euclidean (L2) and the Manhattan (L1) distance measures leads to similar learning results. For this reason, the L1 measure has been implemented, as in this case the circuit is much simpler than the one using the L2 distance, resulting in very low chip area and low power dissipation. This enables including even large NNs in miniaturized portable devices, such as sensors in Wireless Sensor Networks (WSN) or Wireless Body Area Networks (WBAN).
The authors describe a technique for the automatic acquisition of expert knowledge in order to set up a knowledge base for the diagnostic classification of ECG signals. The method is indirect, because the knowledge of the expert, in contrast with the general approach which learns through the direct communication of rules and facts, is derived from a learning set of classified ECGs. It is, on the other hand, different from conventional statistical techniques, because (1) the reference classification is given by experts and not by independent exams like autopsy, coronarography, echocardiography, cardiac surgery, and so on, and (2) this classification can be uncertain, i.e. the various classes are associated with each ECG with certainty factors which can differ from 0 or 1. The data are derived from the CSE pilot diagnostic library. In this preliminary study, the results of the method, which is based on fuzzy pattern matching, show a global type-4 error (complete disagreement) equal to 12.5%.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>
Energy saving becomes a central issue in the design of wireless sensor network routing algorithms. In the wireless sensor networks (WSNs), when intra-network communication is ensured, the lifetime of node can be extended by reducing data transmission or data volume as much as possible. However, the problem is that energy of the nodes around the sink node becomes exhausted quickly due to excessive communication overhead. To handle this problem, in this study, we propose a routing algorithm based on the sink node path optimisation. The study uses the energy consumption model as a constraint, transforms the time optimisation problem into an optimisation model, optimises the sink node path with the aid of simulated annealing (SA) algorithm, and uses data fusion to reduce the intra-network redundant data in the time domain. The proposed algorithm innovatively self-adjusts the path of sink node that is optimised by SA using new fitness function. Comprehensive simulation results show that the proposed algorithm can reduce the node energy consumption of waiting of sink node at the address of sink node, balance the network load and improve survival time of WSNs by 30% in comparison with results produced with the state-of-the art algorithms REAC-IN and DALMDT.
